The future of research: Can we go full digital?

Episode 4 of the NIHR podcast series discussing the big research challenges facing healthcare, hosted by Dr Sanjay Ramakrishnan, Research Fellow in Respiratory Medicine at the University of Oxford . Every week, UK clinicians with experience of research discuss the tough questions about research.

In this final episode,  Sanjay is joined by research optometrist Laura Taylor, ophthalmology registrar Dr. Mital Shah,  and Dr. Tanya Baron, an emergency medicine consultant and former GP, to discuss to what extent health research has kept up with the wider digital revolution taking place in healthcare.

The views and opinions expressed in this podcast are those of the host and guests and do not necessarily reflect those of the NIHR or the Department of Health and Social Care.

Transcript

Sanjay Ramakrishnan  00:00
Welcome to this special podcast series from the National Institute for Health and Care Research about the future of research. You'll be hearing from five clinicians who are the current Thames Valley and South Midlands Clinical Research Network fellows. We will be discussing the big research challenges facing healthcare. I'm Dr. Sanjay Ramakrishnan. I'm a research fellow in respiratory Medicine at the University of Oxford. And over four episodes, I'll be asking my colleagues the tough questions. Are we getting trial participants who are representative of our diverse population? Is our research workforce ready for the future? Is it all going to be online? Join us to get answers to all these tough questions. The title of the podcast today is the future of research. Can we go full digital? The digital revolution is happening in healthcare, is research there? What opportunities and challenges does this provide? To discuss this topic? I'm joined by Mital, Laura and Tanya. To start with can I just ask each of you to introduce yourself to our listeners.             

Laura Taylor  01:05
Hi, my name is Dr. Mital. Shah and I'm an ophthalmology registrar and I work at the Oxford eye hospital. And over the last six years, I've been involved in the running and delivery of a number of different clinical trials. And my research focuses on the use of retinal imaging and artificial intelligence in ophthalmology.

Mital Shah  01:22
Hi, my name is Dr. Mittal Shah and when I was managing registrar and I work at the Oxford Eye Hospital. Over the last six years, I've been involved in the running and delivery over a number of different clinical trials. And my research focuses on the use of retinal imaging and artificial intelligence in ophthalmology.

Tanya Baron  01:40
Hi, I'm Dr. Tanya Baron. I'm an emergency medicine consultant in Oxford and have previously been a GP. I'm involved in various trials within the emergency research and chair, our local patient and public involvement group. My passion is to make research accessible to everybody and to ensure is a normal part of patient care.

Sanjay Ramakrishnan  01:59
Thank you all for joining, thank you team. So we I think it's fair to say health research has not kept up with the wider digital revolution that is already happening, I think, even, you know, in wider healthcare space, how much of it's done electronically is probably quite far ahead of what is done in day to day research practice. It starts from recruitment to consent to data collection to follow up and dissemination. We think it's there if the patients expected it, they're always surprised by how much paper we use. So I guess we can start from scratch. When you join a study, your participant signs a consent form. And that's where it starts, literally a piece of paper as a wet signature on a piece of paper. And there are studies now that are moving quite quickly towards e-consent again, pushed quite quickly by COVID. And, Tanya, you'd have a lot of experience with this. How does it work for you? How does your department handle electronic consent?

Tanya Baron  03:04
So I think within the emergency department, we had started to move towards electronic consent before COVID. But COVID really accelerated things. So the trials all became electronic consent. And it's worked really well for us. It saves you having lots of pieces of paper everywhere. And it's obviously greener for the environment. And actually patients have responded well to this, the research nurses obviously still come and talk to the patient, explain the trial, but the patient information sheet will probably be online, but also paper available. And then the patient can go through the consent form electronically, on an iPad, usually. And yeah, we've had no problem at all, we're able to clean the iPads between each patient. And it's, as I say, been great to be able to enter the information directly onto the platform that we're using for the study data as well.

Sanjay Ramakrishnan  04:09
There's all sorts of hybrid levels of this, the everything from you still do a standard conversation. And only the last bit instead of there being an actual signature you're doing it online. Or the good example is the COVID-19 vaccine study where groups have to get patients through as quickly as possible groups of volunteers we've got come in sit in a room socially distance watch a video recording of someone explaining the vaccine study. And then if they had any questions they will ask in that group setting to someone in the front is almost in a classroom, and everyone gets the answered questions commonly and everyone signs their  consent form electronically at the end of that of that session, again, 45 minute session for consent for a group of people. And then the final step is everything online. You read your information sheet online, you watch videos online, can call and ask questions if you like and you sign the consent form online. That's it. That's your consent process done. Obviously very different each stage, is there any issues with that Laura or do you do you think that any particular stage of that could be could be done better is should we or should we all be just doing all consent at home online patients is reading their own videos, or in their own.

Laura Taylor  05:30
I think I think I've got quite strong opinions on this. I'm all for sharing information online or you know through a video or having documents of either all e-mailing information to patients, and even emailing the input the consent form in advance. But in my opinion, I think there needs to be for Informed Consent truly informed consent for any condition or any procedure or research, I think there needs to be face to face or a telephone back and forth conversation with a human being not a computer, where there is understanding from both sides that you know, you understand what you're agreeing to. And you have the opportunity to openly ask questions, if you've got a silly question, you might not feel like you could e-mail it, if it's done. If you're e-mailing a consent and filling it out on your own at home, you might not feel you could ask a silly question. Whereas if you have the opportunity to have a conversation with someone, also, when I'm consenting someone, the idea is you, you make sure that the patient or the participant understands and relays back to you what they're consenting to. And I don't think you can do that electronically, without engaging with the patient. So I think there's a scope for it to use digital resources in consenting, I think it can speed up the process, I think it can make information a lot more accessible. In the Eye Hospital, if we had more information available online or in electronic resources, you know, we could make it bigger for visually impaired patients, they could use their, you know, they have special different computer devices at home to magnify things up to make it clearer to read. I think that's great. But I still think there needs to be a face to face or telephone conversation.

Mital Shah  07:29
I disagree slightly, with what Laura said. And I think, personally, I'm a huge proponent of moving things forward and into a sort of a digital context. I think above all, it's important to remember that there's always going to be certain groups of people who may not be either digitally literate or savvy or potential access to digital resources in order to be able to take part in that. So I think there should always be a provision for those people so that not excluded from potential research projects going forward. But I think the whole process of consent is just that it's not a signature on a form be that a wet signature on a bit of paper or electronic signature on electronic form, it's a process and I think, taking that into consideration, it will depend on the study that's in question. If you've got something that's an interventional study that says delivering a new treatment, and that treatment is to be given by an invasive procedure, your whole consent process is going to be completely different to a small qualitative study, which just uses a short really short questionnaire, which may take two minutes to fill in to answer a different research question. And in that latter example, where it's just a few questions to answer, if it's really straightforward, the participant information sheet surely will explain everything there enough detail. And as long as there's always an option for people to get in touch, it may not actually require anybody to pick up the phone and call them to check the patient's Okay, or the participants, okay? Or otherwise, we'll all even have a face to face discussion. It may be something that can be done entirely online. But I understand if it was something more invasive, you'd want to have more opportunities to check for the positive and understand what's going on, and to give them an opportunity to ask questions going forward as well. So I think it's a process and I think I definitely wouldn't let this prevent or prevent sort of us progressing towards that direction.

Sanjay Ramakrishnan  09:31
I guess, digging more aside on this, it's all a matter of degrees, isn't it Mital? It's a matter of drawing a line saying this is okay to not require a phone call. But this does and in degree of invasiveness is different for everyone. Even the four of us on here will decide different things that are not invasive. You know, I'm a chest physician, I put cameras into people's lungs. And I don't think that's I think that's minimally invasive. But my patient said, you are absolutely out of your mind that's absolutely invasive. So it's a line that you draw who who draws that line?

Mital Shah  10:09
I think that's a line that's drawn, taking everyone's account or everyone's opinion into account. So researchers are setting up the study, of course, the participants, or like to ppi, patient and participant in sort of groups, and also the research ethical committee and the checks and balances that go in before a study is even given the green light to go ahead. So  you're trying to make sure you get a wide ranging opinion on whether it's considered to be appropriate or invasive. are not enough or to make sure that your consent process sort of stacks up to your study that you're trying to set up.

Laura Taylor  10:53
I think when you're consenting someone in person, I mean, again, it does depend on the level of invasiveness of the study or the treatment. But you kind of judge that person's behaviour, you know, their persona, are they emotionally are they in a good place to take part in this study? Or are they the good place to the treatment, and you can't get that judgement or gain that judgement, if it's done on a digital form online, where you don't see that person or you don't speak to that person. And I think it loses the human aspect, human aspect of it. And I think that's sad. 

Tanya Baron  11:35
I think it was very much dependent on the trial, as Mital was saying, as well. I'm doing a study just now actually, in the emergency department where it's all just been online by email for consent. And I think as long as the researcher is accessible and easily accessible, then I think that's not unreasonable. 

Sanjay Ramakrishnan  11:55
So the next step after consent, all research is about data and how we collect data. And again, the vast majority of time there is one database that collects data that is blood tests, or, or X ray, something else, how patient feels blood pressure, and there's another database, it's a secure database that we then transfer the information, this case to the next one. And that in itself has its own training programme, we sometimes call it good clinical practice training. But that's a whole training that goes into that. But that step doesn't need to exist. And one of the great things that the NHS provides researchers, something we call NHS digi trials. So for things that are very, very clear cut your either alive or you're you've died, unfortunately, or you you either have cancer, that's progressed, or cancer that hasn't progressed, you either have pneumonia or not. Why don't we just see if someone presents the hospital with something to consider that was the outcome that you're testing? That's NHS digi trials that now provides this information to researchers after they have ethical approval to say, if a patient's already agreed, can they, can we just not see them again, they don't have to be followed up, there's less burden on the patients. We just wait to see if the event happens the event in question that we're worried about happens. And then the NHS digi trials process will provide the information to the study. Mital, I think I already know where you're going to stand on this.

Mital Shah  13:32
Yeah. So again, I think I'm a huge proponent. So the NHS has one, I mean, it's got lots of valuable resources. But one resource in particular, that's extremely valuable is all the data that it holds. I think on a practical day to day level, sometimes it can be difficult, because sometimes these datasets are sort of siloed in different areas within the NHS or different programmes. But I think something like this, where you're able to pull that data, it's able to be sent off securely and anonymized or pseudonymized, to researchers who've got approved research projects going on, or potentially even NHS digital trials can be used by potential researchers with approved research studies to try to identify where their potential participants might be. So they can work out which areas of the country for example, might be good research sites, because this condition that they're interested in looking into, it happens to be more common in a particular part of the country. But this has benefits both for researchers because you're less likely to lose participants over time, because they've moved out of the country or not other countries, or if they've moved out of the area, etc. They change GP practice. It also means it's less burden on the participant having to come back in if this is data that's being collected anyway, why are we then asking participants to come back to a hospital or another base where we can ask these questions, again, where the data is actually being collected on its own already? So yeah, I'm a proponent.

Sanjay Ramakrishnan  15:02
We lose the gratitude that we are deeply indebted to these trial participants? Are we still providing the extra service that we should be providing?

Laura Taylor  15:15
I think it depends what the research studies for I mean, I think NHS digi trials, correct me if I'm wrong, but I think it lends itself to sort of big data studies, you know, looking at is it looking at overall big outcomes or big epidemiological studies? And from that aspect, I think it certainly has a place. But I think if you've given a patient a new intervention, I think it's I think you'll lose your potentially risk losing information if you're not going to follow them up, you know, yes, you if you're giving patients you've got two groups, one patient takes aspirin for pneumonia following pneumonia, the patient doesn't. But just by looking at NHS digi trials to see which patients have the longest life expectancy following that, how would you know which patients had better quality of life, that data is not collected? And so potentially you run the risk of missing out on valuable information.

Sanjay Ramakrishnan  16:13
Tanya, Laura touched on the the magic magic word, the quality data, and you're again in the perfect person to talk about this. One classic example in respiratory diseases is, what is the difference between bronchitis and pneumonia. And you deal with this, everyday it's the most common data problem. If it's possible that someone is incorrectly labelled pneumonia, especially if the person seeing them is not senior enough or not experienced enough, when they just had bronchitis or the other way around, they might have had pneumonia, or they might have had pneumonia, but they were incorrectly labelled as bronchitis. And that's it, there's no way of checking. How does that impact your your data collection for your studies?

Tanya Baron  17:00
As I was saying, before, we've got such a huge wealth of data coming in, through the department every day, we have patients with all sorts of things and a lot of patient contact. And I think one thing that is being worked on by the College of Emergency Medicine, is trying to improve the collection of data because we have such as I say, a wealth of data, but actually the systems for collecting it and then for analysing afterwards are, are not have not been great historically. And so if somebody comes in with one particular issue, you could end up being coded something completely different. And then when you're trying to search the data in the future it's very difficult to know whether you're accurately representing what you're what you're looking for. But I mean, there's identified as a problem and is being worked hard on so hopefully that will improve.

Mital Shah  17:59
Yeah, so I think I agree with what sort of Tanya was saying there about that labelling of data, your example that you give Sanjay or this is bronchitis, or is this pneumonia? That's routinely collected data. So that's data that's been collected as part of that patient's normal NHS care that they would have sought for, or that there was some sort of intervention or help for. And I think it's important researchers who are looking to use that kind of data need to and I'm sure we'll understand that it comes with certain limitations. But the potential benefits there are you've got studies, which that data is still able to answer certain questions, it's not going to be able to answer all questions. And Laura mentioned before, but if you've got a new study, and you want to look at health related quality of life versus one treatment versus another treatment, fine that data is not necessarily going to be able to give you that. But I think it's related to what your research question is. So I think there's still huge scope for using the data that's there and an already accessible there. It will make studies potentially complete more efficiently than they would otherwise mean reduced costs related research, if you didn't have to get paid participants to come back in again, for participants and their point of view. If it's a case of yes, I'm happy for you to use this data, which has been collected anyway, for this important research question. They can be involved in the study, but again, from their point of view, is less inconvenience for them had to have to then be followed up and things like that as well.

Sanjay Ramakrishnan  19:29
Yeah, and a lot of studies are moving that way. Partly, some of them had to for the pandemic, but now people realise that people could fill out questionnaires, patient reported outcomes, quality of life data, doesn't even need to involve actually physically coming to a clinic. It can be done from home. And that is another great change that has happened a lot due to the pandemic. But moving all the way back to recruitment, sort of touched on using the digital data to guide to where to go. But one of the big things that the NIHR is doing is setting up the be part of research database, and it's encouraging members of the public to go and sign themselves up to be contacted, when the study opens that's relevant to them. And also self identify studies to join, can we leverage the NHS data even more can can they send the patient studies when something opens based on routinely collected data? So you get a notification on your app saying, hey, there's a new study that's open locally. Are you interested? Is that okay? Is that too invasive?

Laura Taylor  20:42
Um, I think that would be okay. I mean, I wouldn't have a problem every time I log into the NHS app, I get a ping or a notification to say, oh, there's a study in your area that could be relevant to you, you've got if you've got asthma, you know, there's a new asthma study in your area, this might be relevant to you click here for more information, I think, you know, I might not have gone out of my way to look for it. But if it pops into my inbox, or say on the app on my phone, it might prompt me to do that. And I think I don't see that as a as a bad thing. Certainly not. I think we're used to having notifications and pop ups these days on our phones.

Mital Shah  21:25
I think I'd agree again, there with what Laura said, I think the one thing to bear in mind is if someone has opted out, for their data to be shared, sort of off at that level that's from NHS digital, or from the spine, then there's potentially going to be a pool of participants who may not be notified about research studies, where they may have opted out for the data not to be shared, but not necessarily have realised that the implications and things like that could have had as well. So that’s potentially important to bear in mind as well.

Tanya Baron  22:02
Yeah, I think I like the idea of making it as easily accessible as possible for people to be involved in research so that it's very quick and easy. And if they just have to click on yes, I'll be involved in that. And hey presto they're off. But I think again, we have to be mindful that one of the big goals is to ensure that we're inclusive, inclusive, and actually, are we putting more barriers up? By doing it that way? Are we going to bias the participants again, because they're going to be people who are sort of proactively sort of involved in research? Or? Or is it just another way of getting more people involved? And does it actually matter?

Laura Taylor  22:46
I think I disagree. Because there's more people these days that have access to computers and smartphones, and especially in the UK, I think we're quite lucky, quite a lot of people have digital access. And it was more people that do than don't. And I think as time goes on, you know, I think that group of people that do not have access is going to get smaller and smaller.

Tanya Baron  23:10
There are still going to be the people in the groups who are trying to already include more that are probably going to miss out.

Mital Shah  23:20
I think there's always going to be small areas of society who may not have access to the internet, or computers or smartphones. So I think, whatever research study, and no matter how far this sort of digital push goes, and there always has to be a fallback, one to make sure that those participants aren't all potential participants aren't excluded. And two, if a participant comes along, who is able to use it, but says, You know what, I want to see a piece of paper and I want to come in and talk to you. And I don't mind coming in and talking to you and signing a consent form with a pen rather than electronically. I think there always should be an option available for those if that's what they would like.

Laura Taylor  23:57
Moving slightly on, I've got a different question that we've not discussed. And I was wondering what your opinions, I had a conversation recently with someone and they were talking about how you know, your your smartphone data that tracks your everyday life, everyday movements, or how your internet usage, what you browse, or how big you have the font on your phone, or the brightness on your phone can be used with artificial intelligence to monitor for certain health conditions and can then be used to send a warning of perhaps, you know, low mood or a certain health condition and recommend you get help or GP or I just wondered, is that something we should? I mean, it falls into digital in slightly into research, because it's a new thing. What do you guys think about is that something that we should? Is that something we should be embracing?  

Sanjay Ramakrishnan  24:52
Yeah, I think that's a great point. Actually, that's very relevant to our discussion. It is routinely collected data, right? It is data that we're collecting every day every second everyone else uses it, Amazon uses it to target what to sell us. Why can't we use it for research? Why if we can pick up macular degeneration earlier by doing that, if we pick up depression earlier based on the kind of videos you engage with on social media, what is it? You know, where do we draw the line where we go from healthcare to minority report, you know, where where do we sit on that, that on that Continium? What about that what do you think about that Tanya?

Tanya Baron  25:36
I don't know. I have to say it sounds pretty intrusive. I've already had adverts for the the podcast recording company on my phone this lunchtime. And if somebody then says I started doing a pop up saying you might be depressed because you've looked at this thing, I think I'd probably be slightly concerned. I don't know, it seems like it maybe overstepping what we're used to. But  maybe that is the way to go.

Sanjay Ramakrishnan  26:04
Mital is the AI researcher here. And he knows more about what's happening in this field, especially deep mine AI doing a lot of work in this field, what is going on in this field?

Mital Shah  26:17
So, whether or not an ophthalmology context, so that, certainly within ophthalmology, there's lots of interest in using the data that we collect. And one big thing in Ophthalmology is retinal imaging. So that's images of taken of the back of the eye. And trying to use automated means through artificial intelligence and machine learning. To be able to diagnose people automatically without the need for an ophthalmologist or an optometrist or any other healthcare practitioner, and potentially even trying to be able to use that information to predict when someone might need treatment or, or one thing is may change, which is something that we can't do at the moment. So I think there's a lot of interest in the area, I think there's a lot of promise in the area and people are hopeful that big things will come of it. And I think as with anything else, the research will tell us what works and what doesn't work. I mean, there are already or there's already at the very least one AI driven programme, which was approved in the USA for automatically grading or monitoring or identifying I should say diabetic retinopathy from from retinal images. So I think time will tell I think more stuff is coming out on the site, lots of research within the area, and how that fits into our day to day clinical work. We will see.

Sanjay Ramakrishnan  27:40
On that note, I think we can end on that hopeful optimistic note researchers are nothing if not optimistic, to say that there is so much potential in this data world and and the human side of data, you know what, what having this digital access allows the humans involved to achieve and how it changes the experience for participants. There's so much that can be gained by going digital in research and I think today we prove that there's just an unlimited stream of opportunity. Thank you to Laura, Tanya and Mital, so much for joining and thank you to the listeners. Thank you for listening to this special podcast series from the NIHR. If you're a member of the public interested in research, please visit the Be Part of Research website to search for studies near you that you can take part in. For health workers who want to find out more please visit the NIHR Your Part in Research website to get started with your specific research, training and research career related questions please speak to your local friendly NIHR Clinical Research Network.